Improving Resource and Energy Efficiency for Cloud 3D through Excessive Rendering Reduction

Date

2024-04-22

Authors

Liu, Tianyi
Lucas, Jerry
He, Sen
Liu, Tongping
Wang, Xiaoyin
Wang, Wei

Journal Title

Journal ISSN

Volume Title

Publisher

Association for Computing Machinery

Abstract

The rise of cloud gaming makes interactive 3D applications an emerging type of data center workload. However, the excessive rendering in current cloud 3D systems leads to large gaps between the cloud and client frame rates (FPS, frames per second), thus wasting resources and power. Although FPS regulation can remove excessive rendering, due to the highly-varying frame processing time and the use of rendering delays, existing cloud FPS regulation solutions have low FPS and slow motion-to-photon (MtP) latency, causing violations of Quality-of-Service (QoS) requirements.

In this paper, we present a novel cloud FPS regulation solution, called OnDemand Rendering (ODR). ODR employs multi-buffering, dynamic rendering delay/acceleration, and input processing prioritization to reduce excessive rendering and ensure QoS satisfaction. ODR was evaluated in our private cloud and Google cloud. Evaluation results showed that ODR effectively removed excessive rendering, thus improving DRAM performance by 19% and reducing power usage by 16% over no FPS regulation. Better memory efficiency also allowed ODR to increase client FPS by 5.5%. Moreover, ODR reduced average MtP latency by more than 92% and outperformed existing FPS regulations. More importantly, ODR's high FPS and low latency make it feasible to deploy 3D applications to conventional public clouds.

Description

Keywords

cloud graphics rendering, resource and energy efficiency, FPS gaps, OnDemand Rendering, priority frames

Citation

Liu, T., Lucas, J., He, S., Liu, T., Wang, X., & Wang, W. (2024). Improving Resource and Energy Efficiency for Cloud 3D through Excessive Rendering Reduction. Paper presented at the Nineteenth European Conference on Computer Systems, Athens, Greece. https://doi.org/10.1145/3627703.3650064

Department

Computer Science